http://www.rohm.com/products/databook/led/pdf/spr-54.pdfHi Mike,
NJ beat me to it but I agree that the battery that you used was a little light for the job. A Fluke on diode check is a great way to check LEDs. You'll find that it takes slightly more voltage to bias green Vs red as shown in the above link. That is also reflected in observed meter readings. As long as you are sure of LED polarity, you can use a 9V transistor radio battery and about 700 ohms to establish around 10mA as a safe test current. But 9 volts could exceed the reverse breakdown voltage on an LED that you are not sure of. In that case, better to use 2 of your dry cells in series with about 100 ohms as a safe test setup. I haven't noticed an LED with < 5 volts reverse breakdown voltage. In short, take the voltage source you have, subtract 2V for the LED. Divide the remaining voltage by the biasing current that you desire (10mA is good) to get the ballpark resistor value. Reds or greens will boldly reveal their true identity, while whites or blues might seem a liitle shy with their higher forward voltage. (Note: 3 volts wouldn't be enough to check blues or whites)
Joe
Of course, now I see that the datasheet I linked to shows a max reverse voltage of 3V. Also, here is a link to 1, 10mm Red/Green LED
http://media.digikey.com/pdf/Data%20Sheets/Lumex%20PDFs/Ssl-lx100139IGW.pdf
